Business Intelligence (BI) is a catch-all term for a wide variety of software, covering a broad spectrum of information technologies: data collection and transformation, data storage (databases), information management, reporting, decision support, analytics…

The Big Data phenomenon is addressed by a subset of BI technologies – many of them highly specialised to Big Data needs, but as Big Data evolves, many vendors are adding Big Data capabilities to existing software suites.

Retailers have embraced BI for many years; indeed many BI vendors developed their first products for clients in the retail sector. As such, retail has been a fertile ground for Big Data thinking – Amazon and eBay being two of the poster children for Big Data and data-driven business philosophy.

Most BI implementations, however, have focused on reporting on “known knowns” – that is, most retailers have implemented BI technologies to enable them to monitor expected performance. Tens of millions of pounds have been spent creating giant databases, from which data is drawn to be compared against expected norms.

Budgets, targets and forecasts are prepared, and then actual performance compared against them. Most of the retail sector is obsessed with a relatively small number of specific known measures; year-on-year comparisons, category average margin, week-on-week growth/decline, average basket value etc.

More recently, retailers are recognising that the devil is in the detail (they do say “retail is detail”) and that averages gloss over and obscure true patterns of performance. Averages mask the extremes of good and bad performance; the average performance for all products in a category may be fine… but it will mask exceptional performance from some products, and awful performance from others.

Similarly, analysing performance by store, shelf, time of day and individual shopper highlights many fascinating sub-groups – and understanding these is the key to success in a hostile marketplace.

Analytics promotes the exploration of “known unknowns”; those questions about the business which can be posed easily enough, but are difficult to answer with certainty:

  • “Which promotional mechanics work best for category X?”
  • “What is the minimum credible range I can provide for category Y?”
  • “What are the shopper missions in my convenience stores, and which promotes loyalty?”
  • “How do shoppers interpret my price points, do they perceive them as consistent?”
  • “Why has availability fallen despite improved supplier service?”

These are challenging questions, but they can be addressed through a culture of analytics. Statistical and algorithmic analytical techniques have their place, and much attention is focused on the expanding world of predictive analytics – everyone wants to predict future performance, and analytics is seen as a crystal ball.

Experience suggests, however, that visual analytics – the discipline of converting large data sets into interactive graphical presentations – offers the fastest route to exploration of “known unknowns”. Visual analytics harnesses the power of human visual perception, and enables analysts and executives alike to explore information in an intuitive manner, free from technical jargon or reliance on statistical terminology.